Bladder cancer resulted in an estimated 10,700 deaths in the United States in 1983. Our randomized prospectively-controlled evaluation of intravesical and percutaneous BCG immunotherapy in 94 patients with bladder cancer has demonstrated marked reduction in the rate of tumor recurrence (p=0.0017). These results have now been confirmed by two independent prospctively-controlled studies. While BCG is now accepted as a clearly effective agent in the treatment of superficial bladder cancer, its role in invasive bladder cancer, as well as the optimum dose, route, and frequency of administration remain undetermined. Surprisingly, a recent controlled clinical trial from Brazil has demonstrated that BCG administered orally is highly effective (P=0.001). Oral BCG is potentially superior to intravesically administered BCG. Oral administration has been remarkably free of complications and has the advantage of being potentially effective against tumor that is inaccessible by local administration. Therefore, we propose to compare our current optimal BCG immunotherapy protocol, i.e. intravesical and percutaneous maintenance BCG, with high-dose oral BCG immunotherapy. Moreover, since systemic BCG immunotherapy administered orally has been demonstrated to markedly reduce local tumor recurrence within the bladder, and BCG immunotherapy has been demonstrated to have a complete respnse rate of 56% in patients withe existing superficial transitional cell carcinoma, we propose to evaluate the exciting potential that BCG immunotherapy may be an effective adjuvant in treating the more than 50% of patients who die following cystectomy because of metastatic transitional cell carcinoma which was undetectable at the time of surgery. We have demonstrated that in vivo assay of cellular immunity to BCG as determined by the intermediate strength PPD skin test is highly predictive of antitumor response to BCG immunotherapy (p=0.017, Savage). Natural killer (NK) cell activity has been determined to vary with route of administration of BCG and correlate with effective antitumor immunity. We therefore propose to determine NK cell numbers using the fluorescence activated cell sorted; NK cell activity using 51Cr release from K562 cells; and NK cell recycling capacity using the thin layer agarose assay. These parameters of the immune response will be determined prior to the initiation of immunotherapy, during the course of immunotherapy, and at the time of disease recurrence or progression and correlated with the clinical status.